https://github.com/hsm207/grab-safety

My submission for Grab AI for S.E.A. challenge
https://github.com/hsm207/grab-safety

databricks databricks-notebooks pyspark spark-ml telematics

Last synced: 12 months ago
JSON representation

My submission for Grab AI for S.E.A. challenge

• Host: GitHub
• URL: https://github.com/hsm207/grab-safety
• Owner: hsm207
• License: gpl-3.0
• Created: 2019-06-03T15:31:45.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2019-06-03T15:31:53.000Z (about 7 years ago)
• Last Synced: 2025-06-01T12:45:59.749Z (about 1 year ago)
• Topics: databricks, databricks-notebooks, pyspark, spark-ml, telematics
• Language: Jupyter Notebook
• Size: 649 KB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Introduction

This repository contains my solution to the Grab  AI for S.E.A. [Safety Challenge](https://www.aiforsea.com/safety).

# Prerequisites

The [notebooks](/notebooks) were developed on Databricks using Python 3 and PySpark running on the  5.4 ML

Beta (includes Apache Spark 2.4.0, Scala 2.11) runtime.

# Usage

Execute the notebooks in the following order and modify any directory and file names accordingly:

* [01_data.ipynb](/notebooks/01_data.ipynb): This notebook downloads the dataset and saves it in parquet format

* [02_EDA.ipynb](/notebooks/02_EDA.ipynb): This notebook does some exploratory data analysis and removes outliers

* [03_feature_engineering.ipynb](/notebooks/03_feature_engineering.ipynb): This notebook computes features that will be used to build

models to classify a trip

* [04_modelling.ipynb](/notebooks/04_modelling.ipynb): This notebook builds and validates a few models

* [05_evaluation.ipynb](/notebooks/05_evaluation.ipynb): This notebook is meant to evaluate the best model found in the previous

notebook against the hold-out dataset.

More details are available in each notebook.